Title :
Fixed-point roundoff error analysis of large feedforward neural networks
Author :
Choi, H. ; Burleson, W.P. ; Phatak, D.S.
Author_Institution :
Dept. of Electr. & Comput. Eng., Massachusetts Univ., Amherst, MA, USA
Abstract :
Digital implementations of neural nets must consider finite wordlength effects. For large sized nets, it is particularly important to investigate the roundoff errors in order to realize low-cost hardware implementations while satisfying precision constraints. This paper presents output error expressions for a large feedforward neural net, which are based on statistical error analysis. Weight quantization errors as well as arithmetic errors due to rounding of multiplier output and sigmoid output are modeled. The results indicate that for equal wordlengths, multiplier roundoff errors exceed weight quantization errors by about an order of magnitude.
Keywords :
error analysis; feedforward neural nets; roundoff errors; statistical analysis; arithmetic errors; feedforward neural networks; finite wordlength effects; fixed-point roundoff error analysis; output error; sigmoid output; statistical error analysis; weight quantization errors; Aggregates; Arithmetic; Error analysis; Feedforward neural networks; Hardware; Multi-layer neural network; Neural networks; Nonhomogeneous media; Quantization; Roundoff errors;
Conference_Titel :
Neural Networks, 1993. IJCNN '93-Nagoya. Proceedings of 1993 International Joint Conference on
Print_ISBN :
0-7803-1421-2
DOI :
10.1109/IJCNN.1993.717037
